The world of telematics has moved on quite a bit since the days of black boxes being fitted to hot hatches. Driver behaviour stats can be overlaid with precise location, regular routes driven, roadworks, plus the wealth of sensor and camera data being gathered by modern vehicles.
Most of this can be analysed and crunched in real time too, which offers insurance brands the chance to offer PAYG pricing per mile, bespoke commercial fleet products and much more.
IE caught up Charles Morriston, Head of Professional Services and Pre-Sales at VisionTrack, to learn more.
IE; What are the main gains from in-car sensor data, compared to traditional black boxes?
CM; Older telematics boxes offered some useful insights on location, speed, time of day etc. But modern cars have so much data within them and it’s especially important that its accessible after an incident, like a collision, theft attempt and so on.
The spread of cameras also adds a huge amount of context to the old types of data gained from black boxes. Devices like 360 degree connected cameras are adding stacks of data, in real time, which is great for insurers. The downside of that is that there’s no way that human review of the footage can be undertaken, you need AI to look for potential incidents, or separate real collisions from things like speed bumps, poor parking and so on.
We have also found at VisionTrack that separating this data also provides useful content for driver training across the commercial sector. Effectively, you can begin to ID risk scenarios and play the footage, slow it down, look at it from different angles, various data sources etc. and this gives you insights into how to avoid those higher risk situations on the road.
By flagging NARA incidents (Notification, Analysis, Risk Assessment) as “real” or false positives, then you begin to narrow down the true risk across the fleet; routes, load content, vehicles, parking locations and more.
IE; What effect has the recent expansion of cycle lanes and LTNs made to the commercial delivery sector?
CM; It’s made a big difference. Moreover, the Highway Code changes that give vulnerable road users having priority also means we need more data from vehicles in terms of proximity alerts, sensors, blind spot cameras and more. If you just look at London, HGVs make up 3% of road traffic, but account for the majority of fatal cyclist accidents, which thankfully are low each year, but it highlights the blind spot problem.
AI sits inside the vehicle now, which means it detects when a pedestrian, cyclist or motorcyclist enters the blind spot. The driver gets a visual and audible alert immediately. VisionTrack has done a long term study in NYC with a partner company and learned a great deal from large truck deliveries and the tech used to monitor them. There are lots of safety and training improvements that can be made and it helps with mitigation of incident causation.
IE; Do you think that things like driver score league tables and rewards can help engagement when it comes to fleet telematics?
CM; I’ve been in telematics a long time and carrot always works better than stick. The best way to use the increasing amounts of sensor and camera data is as a training tool, not apportioning blame or singling people out. Look for patterns of behaviour, not individual errors. If you make scorecards anonymous then you can ID high risk behaviours without entering a blame game.
IE; Can Cloud based systems help reduce theft or damage claims, or, say, identify particular locations or routes which have potentially higher risks associated with them?
CM; That’s an interesting one. You can examine data at a very detailed level; vehicle doors being open by location, by time, which days of the week are more likely to see damage or theft. The thing about VisionTrack and all Cloud or Edge systems is that you can develop a customised platform, a dashboard, where your fleet manager can see the data they want quickly, plus how it links to other parameters. For example, shift patterns vs incidents.
IE; How about the future, will there be a time when vehicles can recognise other vans or cars following, or using the same location and then match the plates and dates to specific incidents, like crash-for-cash or thefts from vehicles, warehouses etc?
CM; Factoring in ANPR data to systems is something the industry is looking at. It’s on the horizon for sure and although there are privacy issues of course, it’s inevitable that the market will demand more leverage from data. It’s really all about context and patterns, so even anonymised data can help insurers and fleet operators prevent things happening. That’s the big win.
IE; Interesting insights, thank you.
Be the first to comment